DPDK patches and discussions
 help / color / mirror / Atom feed
* Ethertype and Rx/Tx offload suites - collaborate with Luca
@ 2025-03-06 18:52 Patrick Robb
  2025-03-06 22:06 ` Dean Marx
  0 siblings, 1 reply; 2+ messages in thread
From: Patrick Robb @ 2025-03-06 18:52 UTC (permalink / raw)
  To: Dean Marx, Nicholas Pratte; +Cc: Luca Vizzarro, Paul Szczepanek, dev

[-- Attachment #1: Type: text/plain, Size: 1105 bytes --]

Hi Dean/Nick,

Luca said this morning he is starting to look at Ethertype and rx/tx
offload testsuites. At the CI meeting I gave him an overview of what you
both had learned about these so there isn't double work. My summary:

Ethertype:

1.TPID/VLAN filtering works only on Mellanox NICs and we don't know why so
we are blocked pending this ticket:
https://bugs.dpdk.org/show_bug.cgi?id=1464

2. Nick I mentioned how you were looking at Intel's offering of supporting
non-standard TPIDs and how this was not done by other vendors, and raised
the decision of supporting these testcases, or scoping the testsuite down
to only standard TPIDs.

3.If you can share your patch files with Luca and describe where you landed
with the device capability check for this it may prevent double work.

rx/tx offload:

1.Dean has rebased this series but is running into a bug with the
capability check test parser returning no flags. He confirmed from running
the steps manually that the flags we expect are being reported by testpmd.

2. Dean please provide a summary to Luca so there is no double work.

Thanks guys.

[-- Attachment #2: Type: text/html, Size: 1432 bytes --]

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: Ethertype and Rx/Tx offload suites - collaborate with Luca
  2025-03-06 18:52 Ethertype and Rx/Tx offload suites - collaborate with Luca Patrick Robb
@ 2025-03-06 22:06 ` Dean Marx
  0 siblings, 0 replies; 2+ messages in thread
From: Dean Marx @ 2025-03-06 22:06 UTC (permalink / raw)
  To: Patrick Robb; +Cc: Nicholas Pratte, Luca Vizzarro, Paul Szczepanek, dev

> rx/tx offload:
>
> 1.Dean has rebased this series but is running into a bug with the capability check test parser returning no flags. He confirmed from running the steps manually that the flags we expect are being reported by testpmd.
>
> 2. Dean please provide a summary to Luca so there is no double work.

Hi Luca,

Essentially what I ran into while trying to rebase the series was an
issue with the parser in the OffloadCapability class. The make_parser
function contains an argument called "find_multiple", which will match
all port/queue capabilities/configurations found in testpmd, rather
than just the first instance. This argument is set to True and runs
during test suite execution, but causes an error because one of the
variables sent to from_list() during the parser creation is not a list
of strings (like it's supposed to be) but instead the string "0". I
moved away from the series to work on some other items in the
meantime, but I was thinking it could be an issue with the regex that
is grabbing an extra string. In this instance, the command "show port
0 rx_offload configuration" is sent to testpmd, which prints something
like the following:

testpmd> show port 0 rx_offload configuration
Rx Offloading Configuration of port 0 :
  Port :
  Queue[ 0] :

If I had to guess, the "0" string is coming from the last line, but
again I never finished debugging so just an educated guess. If you run
into the same problem let me know.

Cheers

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2025-03-06 22:06 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2025-03-06 18:52 Ethertype and Rx/Tx offload suites - collaborate with Luca Patrick Robb
2025-03-06 22:06 ` Dean Marx

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).